Point cloud representations encapsulate both spatial coordinates and associated feature data including RGB values and surface reflectance properties. Point cloud attribute coding aims to reduce data size by eliminating statistical redundancies in feature values. Point cloud attributes present greater compression challenges than geometry data due to their higher complexity and more intricate redundancy patterns. Contemporary attribute coding approaches utilizing conventional transformation and prediction techniques demonstrate competitive compression performance, while deep learning–based methods are constantly evolving and can have a great potential for optimization. This chapter provides a comprehensive review of the deep learning–based static point cloud attribute coding methods.